[Deploying Sakai] MySQL performance issues with 2.8

Matthew Jones matthew at longsight.com
Wed Dec 5 06:43:59 PST 2012


Nothing was added back to 2.8 and many of those kernel tables will be the
slowest without caching. In 2.9 caching was improved around the kernel
tables and it's possible that you might be able to change some of your
cache parameters specific to your institution to lower the hits on the
other kernel tables.

https://jira.sakaiproject.org/browse/KNL-826 - Improve performance by
precaching Sakai Users
https://jira.sakaiproject.org/browse/KNL-934 - Authz Role Key Cache

I don't know of any significant performance improvements to sitestats other
than in the admin (global sitestats) interface. Steve Swinsburg would
probably know as he's done the most work on that tool recently.

With msgcntr, there were reports about it being slow because certain
versions of MySQL were selecting a slow query plan. (
https://jira.sakaiproject.org/browse/MSGCNTR-575) Earle said he had the
correct indexes, just mysql was still running slower queries. But yea,
you'd have to post more specific information.


On Wed, Dec 5, 2012 at 9:35 AM, Sam Ottenhoff <ottenhoff at longsight.com>wrote:

> Hi JFL,
>
> Sharing the actual queries that are slow will be much more helpful than
> the aggregation you have provided.  For example, based on your MsgCntr
> queries, it's possible we could tell if you are missing some key
> performance improvements committed after 2.8.0 was released.
>
> --Sam
>
>
> On Wed, Dec 5, 2012 at 9:29 AM, Jean-Francois Leveque <
> jean-francois.leveque at upmc.fr> wrote:
>
>> Hi all,
>>
>> I've just checked my most frequent MySQL slow log requests.
>>
>> Apart from the commits that cannot be traced by slow log only, I've
>> found the tables that appear more frequently are from 3 different parts
>> of the code.
>>
>> msgcntr (MFR_*)
>> sitestats (SST_*)
>> kernel (SAKAI_SITE, SAKAI_SITE_USER, SAKAI_SITE_PAGE,
>> CONTENT_RESOURCE_DELETE)
>>
>> Have some of you done some extra indexing or other changes that improved
>> performance with these?
>>
>> Cheers,
>>
>> J-F
>> _______________________________________________
>> production mailing list
>> production at collab.sakaiproject.org
>> http://collab.sakaiproject.org/mailman/listinfo/production
>>
>> TO UNSUBSCRIBE: send email to
>> production-unsubscribe at collab.sakaiproject.org with a subject of
>> "unsubscribe"
>>
>
>
> _______________________________________________
> production mailing list
> production at collab.sakaiproject.org
> http://collab.sakaiproject.org/mailman/listinfo/production
>
> TO UNSUBSCRIBE: send email to
> production-unsubscribe at collab.sakaiproject.org with a subject of
> "unsubscribe"
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://collab.sakaiproject.org/pipermail/production/attachments/20121205/555405b9/attachment.html 


More information about the production mailing list